Article

Structural Constraints of V

accine-Induced Tier-2Autologous HIV Neutralizing Antibodies Targetingthe Receptor-Binding Site

Graphical Abstract

Highlights

d HIV-1 TF Env immunization induced potent Tier-2 neutralizing

antibodies (nAbs)

d Vaccine-elicited nAbs target CD4bs and mimic autologous

nAbs in infected individual

d Autologous nAb-Env complex structure reveals mechanism

of strain-specific neutralization

Bradley et al., 2016, Cell Reports 14, 43–54January 5, 2016 ª2016 The Authorshttp://dx.doi.org/10.1016/j.celrep.2015.12.017

Authors

Todd Bradley, Daniela Fera,

Jinal Bhiman, ..., Sampa Santra,

Stephen C. Harrison, Barton F. Haynes

Correspondencetodd.bradley@duke.edu (T.B.),barton.haynes@duke.edu (B.F.H.)

In Brief

HIV-1 vaccine elicitation of antibodies

against neutralization-resistant (Tier-2)

viruses is rare. Bradley et al. demonstrate

induction of antibodies that can

neutralize a vaccine-matched Tier-2 virus

in a rhesus macaque immunized with HIV

trimers isolated from a HIV-1-infected

individual. Structural analysis revealed

the mechanism of restricted

neutralization breadth.

Accession Numbers

5F6H

5F6I

5F6J

Cell Reports

Article

Structural Constraints of Vaccine-Induced Tier-2Autologous HIV Neutralizing Antibodies Targetingthe Receptor-Binding SiteTodd Bradley,1,11,* Daniela Fera,2,11 Jinal Bhiman,4,10 Leila Eslamizar,5 Xiaozhi Lu,1 Kara Anasti,1 Ruijung Zhang,1

Laura L. Sutherland,1 Richard M. Scearce,1 Cindy M. Bowman,1 Christina Stolarchuk,1 Krissey E. Lloyd,1 Robert Parks,1

Amanda Eaton,1 Andrew Foulger,1 Xiaoyan Nie,1 Salim S. Abdool Karim,6,7 Susan Barnett,8 Garnett Kelsoe,1

Thomas B. Kepler,9 S. Munir Alam,1 David C. Montefiori,1 M. Anthony Moody,1 Hua-Xin Liao,1 Lynn Morris,4,6,10

Sampa Santra,5 Stephen C. Harrison,2,3 and Barton F. Haynes1,*1Duke Human Vaccine Institute, Departments of Medicine, Surgery and Immunology, Duke University School of Medicine, Durham,NC 27710, USA2Laboratory of Molecular Medicine, Boston Children’s Hospital, Harvard Medical School, Boston, MA 02115, USA3Howard Hughes Medical Institute, Boston, MA 02115, USA4National Institute for Communicable Diseases, Johannesburg 2131, South Africa5Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02215, USA6Center for AIDS Program of Research in South Africa, University of KwaZulu-Natal, Durban 4013, South Africa7Columbia University, New York, NY 10032, USA8Novartis Vaccines and Diagnostics, Inc., Cambridge, MA 02139, USA9Boston University, Boston, MA 02118, USA10Faculty of Health Sciences, University of the Witwatersrand, Johannesburg 2131, South Africa11Co-first author*Correspondence: todd.bradley@duke.edu (T.B.), barton.haynes@duke.edu (B.F.H.)

http://dx.doi.org/10.1016/j.celrep.2015.12.017

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

SUMMARY

Antibodies that neutralize autologous transmitted/founder (TF) HIV occur in most HIV-infected individ-uals and can evolve to neutralization breadth. Autol-ogous neutralizing antibodies (nAbs) against neutral-ization-resistant (Tier-2) viruses are rarely induced byvaccination. Whereas broadly neutralizing antibody(bnAb)-HIV-Envelope structures have been defined,the structures of autologous nAbs have not. Here,we show that immunization with TF mutant Envsgp140 oligomers induced high-titer, V5-dependentplasma neutralization for a Tier-2 autologous TFevolved mutant virus. Structural analysis of autolo-gous nAb DH427 revealed binding to V5, demon-strating the source of narrow nAb specificity andexplaining the failure to acquire breadth. Thus, oligo-meric TF Envs can elicit autologous nAbs to Tier-2HIVs, but induction of bnAbs will require targetingof precursors of B cell lineages that can mature toheterologous neutralization.

INTRODUCTION

The HIV-1 envelope protein (Env) is the primary target of neutral-

izing antibodies (nAbs) (Wyatt and Sodroski, 1998; Zhou et al.,

2007). One major obstacle to developing an effective HIV-1 vac-

cine is finding an immunogen that can elicit broadly nAbs (bnAbs)

with the capacity to overcome variability of the virus and to retain

neutralizing activity for most circulating HIV-1 strains (Burton

et al., 2012; Mascola and Haynes, 2013).

Between 3 and 12 months after HIV-1 transmission, most in-

fected individuals develop autologous, strain-specific nAbs to

the transmitted/founder (TF) virus and TF variants (Ariyoshi

et al., 1992; Richman et al., 2003; Wei et al., 2003). The autolo-

gous nAb response drives viral escape and stimulates additional

specificities of nAbs that neutralize escape viruses (Richman

et al., 2003; Wei et al., 2003). This antibody virus co-evolution

persists throughout infection, and in �20% of individuals, it

leads, after years of infection, to development of high levels of

bnAbs (Doria-Rose et al., 2010; Gray et al., 2011; Liao et al.,

2013a; Tomaras et al., 2011; Walker et al., 2011).

Two recent studies mapped the ontogeny of bnAbs and TF

viruses from the time of transmission to bnAb development

and showed that bnAbs arise from autologous nAb B cell

clonal lineages but that only a small number of the autologous

nAb lineages ultimately evolve to neutralization breadth (Do-

ria-Rose et al., 2014; Liao et al., 2013a). Identification of im-

munogens that can induce nAbs against autologous, neutral-

ization-resistant (Tier-2) viruses is a major challenge for HIV

vaccine design, and examples of vaccine-matched, Tier-2

nAb responses elicited by vaccination in primates are few

(Sanders et al., 2015; Willey et al., 2003). Moreover, no vac-

cine-induced Tier-2 nAbs have yet been isolated and charac-

terized, nor have structures of their Env complexes been

determined.

CAP206 is an HIV-infected African individual who later

developed gp41-targeted bnAbs (Gray et al., 2009a; Morris

Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors 43

et al., 2011). As a critical first step in HIV vaccine design, we

sought to map the autologous nAb response and to elicit

Tier-2 nAbs that mimicked this early autologous nAb response

by immunization with Env proteins isolated over the course of

infection. We report here that immunization of rhesus ma-

caques with HIV-1 TF variants from CAP206 induced strain-

specific nAbs to vaccine-matched Tier-2 autologous viruses

in three of six animals. After only two immunizations, one ma-

caque had a high-titer nAb response that targeted the CD4-

binding site (bs) and mimicked the autologous nAb response

observed in CAP206. We isolated a vaccine-induced nAb

clonal lineage (DH427) that potently neutralized the Tier-2

CAP206 6-month virus and recapitulated the observed plasma

neutralization response. A crystal structure of DH427 in com-

plex with HIV Env showed that DH427 bound close to the

CD4bs but also interacted with variable regions in the HIV

Env (loop E and V5-loop), explaining the restricted neutraliza-

tion breadth and the failure of DH427 to evolve to heterologous

neutralization.

RESULTS

Immunization of Rhesus Macaques Elicits Tier-2Autologous NeutralizationWe tracked the evolution of the CAP206 env gene from the TF vi-

rus until 39 months after transmission (Figure 1A). We selected

the CAP206 TF and six additional representative mutant env

genes from 2-, 6-, 12-, 21-, 24-, and 30-month time points and

produced them as recombinant gp140 oligomers that were pre-

dominately trimers (Figure S1A). The antigenic and functional

epitopes expressed on each of the recombinant CAP206 Envs

were determined by SPR assays (Figure S1B). All Envs bound

to CD4 and mAb A32, which binds well to uncleaved trimers,

and the magnitude of CD4 binding increased in Envs isolated

from later CAP206 time points (Figure S1B). The seven Envs

showed binding to a panel of nAbs and bound to 17b, an anti-

body that binds to the CD4-induced conformation of Env, in

the absence of CD4 (Figure S2). Some Envs lacked binding of

bnAbs that target the CD4bs, and others only reacted weakly,

indicating disruption of canonical CD4bs bnAb epitopes in the

CAP206 Envs.

We immunized six rhesus macaques with a cocktail of

all seven CAP206 recombinant gp140 oligomers and collected

plasma and PBMCs before the first immunization and 2 weeks

after each subsequent immunization (Figure 1B). We tested

plasma from Env-immunized animals for the presence of

anti-HIV neutralizing activity using the TZM-bl pseudovirus

assay. Plasmas collected before the first immunization

(week 0) and 2 weeks after the last immunization (week 38)

were tested for neutralization of heterologous, neutralization-

sensitive (Tier-1) isolates MN (clade B) and MW965 (clade

C) and of autologous, Tier-2, vaccine-matched isolates

of the seven CAP206 Envs (Figures 1C and S3A). We

observed potent neutralization titers against the Tier-1 MN

and MW965 viruses in all animals. Low levels of neutralizing

activity against Tier-2 autologous viruses were present in

two animals (Figures 1C and S3B). One animal, rhesus ma-

caque 5173, had potent plasma neutralization activity against

44 Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors

the Tier-2 autologous 6-month CAP206 TF variant virus (Fig-

ures 1C and S3B).

Neutralization of the 6-month virus emerged in 5173 plasma

2 weeks after the second immunization and was boosted by re-

petitive immunization and persisted throughout the remainder of

the immunization regimen (Figure 1D). Plasma neutralization re-

sponses waned 20 months after the last immunization (data not

shown). Thus, immunization with CAP206 Envs effectively eli-

cited potent Tier-1 heterologous virus neutralization in all six an-

imals and low levels of Tier-2 nAbs in two of them; in one animal,

it rapidly induced an early and robust neutralization response to

an autologous Tier-2 virus.

Immunization-Induced Neutralization Mimics EarlyAutologous CAP206 Plasma Neutralization ResponseAutologous Tier-2 neutralization in macaque 5173 was specific

for the CAP206 6-month virus. Env sequences that were isolated

from later time points post-infection from CAP206 contained

numerous mutations that were candidate sites of immune pres-

sure. Two of the very first observed Env sequence changes re-

tained in the 6-month Env were in the V1/V2 and V5 regions

(Figure 2A).

Closer inspection of the V5 sequences revealed extensive di-

versity in this region among longitudinal CAP206 Envs. Specif-

ically, the 6-month Env had a five-amino-acid deletion and

lacked a predicted glycosylation site at position 463 (Figure 2B).

To determine whether 5173 neutralization of the 6-month

CAP206 virus targeted V5, we engineered a mutation in the

6-month virus that introduced a glycosylation site at position

463 (S463N). We found persistent plasma neutralization of the

wild-type 6-month virus for samples across the full immunization

period but no neutralization of the glycosylation site variant

(Figure 2C).

To examine the role of mutations in the V5 loop in early autol-

ogous escape of the infecting virus in CAP206, we tested

CAP206 plasma isolated post-infection for the ability to

neutralize the autologous 2-month and 12-month viral isolate.

Early plasma samples exclusively neutralized the 2-month virus,

and only later plasma samples acquired neutralization activity for

the 12-month virus (Figure 2D). When we reverted the 12-month

V5 region back to the 2-month sequence, the chimeric virus

became sensitive to neutralization by early CAP206 plasma

time points. Reverting mutations in the V1V2 in the 12-month

back to the 2-month did not have any effect on early neutraliza-

tion. Thus, the V5 loop was a target for autologous nAbs in

CAP206 after infection and a V5-dependent autologous nAb

response was elicited in a rhesus macaque by immunization

with a CAP206 TF variant.

Immunization Elicited CD4bs Autologous nAbsWe isolated single Env-specific memory B cells from macaque

5173 PBMC using tetramers of the neutralization-sensitive

CAP206 6-month Env as well as the neutralization-resistant

CAP206 30-month Env. Memory cells (CD20+ and CD27+) that

bound the 6-month, but not the 30-month, Env were sorted

into individual wells of a 96-well plate (Figure 3A). The frequency

of 6-month Env-specific memory B cells was 0.07% (Figure 3A).

Single-cell PCR amplification and transient expression of

0.003CAP206 T/F

1 mo

2 mo

6 mo

12 mo

15 mo

21 mo

24 mo

30 mo

33 mo

39 mo

0

2

6

8

12

14

18

20

24

26

30

32

36

38

Week

Plasma sample

#1 #2 #3 #4 #5 #6 #7

ImmunizationMF59 + CAP206 gp140

(T/F, 2 month, 6 month, 12 month, 21 month, 24 month, 30 month)

Immunization

Tier-1 Tier-2

A B

C

D

050

100150200250300350400450

0 2 8 14 20 26 32 38Week

CAP206 6 month T/F variant virus

ID50

CAP206 T/F and other variant viruses

2 month

6 month

12 month

21 month

24 month

30 month

Animal Week B.MN C.MW965 C.CAP206.2mo C.CAP206.6mo C.CAP206.12mo0 <20 <20 <20 <20 <20

38 990 5499 25 <20 280 <20 <20 <20 <20 <20

38 1036 2125 <20 <20 <200 <20 <20 <20 <20 <20

38 804 1682 <20 <20 <200 <20 <20 <20 <20 <20

38 281 2464 <20 437 <200 <20 <20 <20 <20 <20

38 579 1379 <20 <20 <200 <20 <20 <20 <20 <20

38 400 1663 38 <20 <20

5160

5165

5167

5173

5183

5184

Figure 1. Immunization with T/F Envs Elicits Tier-2 Neutralization

(A) Neighbor-joining phylogenic tree of isolated env sequences from South African HIV-1-infected individual CAP206 from the time of transmission to 39 months

post-infection. Red boxes indicate envs selected for production of recombinant immunogens.

(B) Immunization regimen; six rhesus macaques immunized seven times every 6 weeks with a swarm of seven CAP206 Envs. Plasma samples were collected

2 weeks post-immunization.

(C) Plasma neutralization of heterologous Tier-1 and autologous Tier-2 viruses before immunization (week 0) and 2 weeks after the last immunization (week 38)

measured as the plasma dilution at which relative luminescence units (RLUs) were reduced 50% compared to control wells in the TZM-bl neutralization assay

(yellow, ID50 > 20; orange, ID50 > 200; red, ID50 > 1,000).

(D) Plasma neutralization over the course of immunization for rhesus monkey number 5173 of the CAP206 autologous Tier-2 viruses (T/F, 2-month, 6-month,

12-month, 21-month, 24-month, and 30-month) in the TZM-bl neutralization assay. Dashed line indicates immunization time points.

See also Figures S1–S3.

immunoglobulin (Ig) genes of the sorted memory B cells identi-

fied two mAbs (DH427 and DH428) that specifically bound the

CAP206 6-month Env in ELISA (Figure 3B). As expected from

their sensitivity to changes at position 463, DH427 and DH428

competed for binding of soluble CD4 to CAP206 6-month Env

(Figure 3C). Sequences of the heavy- and light-chain Ig genes

showed that DH427 and DH428 are two members of the same

B cell clonal lineage, which used the rhesus orthologs of the hu-

man VH 3-23 and Vl 2-11 genes (Figures 3D and S4). Members of

the DH427 lineage have short HCDR3s (10 aas); the variable

heavy regions were mutated 4.5% in DH427 and 2.4% in

DH428 (Figure 3D).

DH427 CD4bs Antibody Lineage Neutralized the Tier-2CAP206 6-Month VirusDH427 and DH428 neutralized the 6-month virus but did not

neutralize any of the other autologous or heterologous viruses

tested (Figures 4A and S5). Moreover, DH427 and DH428 did

not neutralize the CAP206 6-month virus with an N463 mutation,

recapitulating the pattern of neutralization observed in the 5173

plasma (Figure 4A).

The CAP206 6-month Env lacked a glycosylation site in

V5; introduction of a glycosylation site in V5 of the CAP206

6-month Env eliminated neutralization by both 5173 plasma

and DH427 lineage antibodies. The CAP206 TF Env also lacked

Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors 45

A B

C

D

Figure 2. Glycosylation of the V5 Loop

Blocks Autologous Tier-2 Neutralization

(A) Highlighter plot showing mutations in the Env

that occur in the longitudinal immunogens from the

T/F sequence.

(B) Amino acid alignment of the CAP206 V5

regions.

(C) Macaque 5173 plasma neutralization of wild-

type and mutant CAP206 6-month viruses over the

course of immunization measured as the plasma

dilution at which RLUs were reduced 50%

compared to control wells in the TZM-bl neutrali-

zation assay.

(D) CAP206 plasma neutralization of autologous

2-month and 12-month viruses compared with

12-month reverted mutant viruses over the course

of infection.

a glycosylation site in V5, but the loop was five amino acid resi-

dues longer (Figure 4B). This virus resisted neutralization by

DH427 and DH428; deletion of the five residues in V5 of the

CAP206 TF virus to match the 6-month virus made it sensitive

to both antibodies (Figure 4C). Thus, both the presence of glyco-

sylation sites and the length of the V5 loop restricted CAP206 vi-

rus neutralization by DH427 and DH428.

Maturation of the DH427 Clonal LineageFor an analysis of the evolution of the DH427 antibody lineage,

we inferred the intermediate (I1) and the unmutated common

ancestor antibodies of DH427 and DH428. The I1 and mature

antibodies specifically and exclusively bound the CAP206

6-month Env, but there was no detectable binding of this partic-

ular unmutated common ancestor candidate antibody to any of

the recombinant CAP206 Envs (Figures 5A and S6). We showed

that light-chain mutations between the unmutated common

ancestor and I1 determined the difference in DH427/DH428

46 Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors

binding by pairing the unmutated com-

mon ancestor heavy chain with the I1 light

chain and vice versa. Only the former had

detectable affinity (Figure 5B). Failure of

the inferred DH427 unmutated common

ancestor to bind the immunogen led us

to consider whether one or more of the

amino acid residues considered to have

come from somatic hypermutation (SHM)

might instead be germline encoded in an

as yet undiscovered rhesus macaque Vl

allelic variant, leading to an incorrect

inference for the unmutated common

ancestor.

We examined the germline IGLV2-F

sequence in macaque 5173 by targeted

genomic amplification and sequencing

of extracted genomic DNA (Figure S7).

Sequencing of individual colonies de-

tected an IGLV2-F allele (IGLV2-F*02) that

shared three of the four residues that we

had considered mutated in I1 (Figure 5C).

We then sequenced the IGLV2-F allele in the five remaining ani-

mals in the immunization study and found IGLV2-F*02 in the

genomic DNA of all five (Figure S6C). From these results, we de-

signed a new unmutated common ancestor (UCA2) by incorpo-

rating the new germline light-chain allele. We found that inferred

UCA2 bound the 6-month Env, indicating that UCA2 was the

authentic initiator of the DH427/DH428 B cell lineage (Figure 5D).

The UCA2 antibody did not neutralize the 6-month virus,

but both DH471 I1 and the chimeric (unmutated common

ancestor VH: I1VL) antibody did neutralize this virus (Fig-

ure 5E). These results indicated that autologous neutralization

activity required a specific sequence of light-chain affinity

maturation.

DH427 Unmutated Common Ancestor Reacts with HostProtein FAM21CTwo macaques (5160 and 5184) had weak autologous nAb

responses, one (5173) had a robust autologous nAb response,

0 102 103 104 105

0

102

103

104

105

gp140CAP206 30 month

gp14

0 CA

P20

6 6

mon

th

0.065 %

A

Log

AUC

CAP206 T/F

CAP206 2m

o

CAP206 6m

o

CAP206 12

mo

CAP206 21

mo

CAP206 24

mo

CAP206 30

mo0

5

10

15DH427DH428

B

C

Antibody Concentration [µg/ml]

%Bl

ocki

ng

1 10 1000

20

40

60

80

100DH427DH428Synagis

sCD

4

Antibody ID

Rh VHgene

(Hu Orth)

HCDR3 Length (AA)

VHMutation

(nt %)

VHMutation (AA %)

Rh VL gene(Hu Orth)

LHCDR3Length (AA)

VLmutation

(nt %)

VLmutation (AA %)

DH427 3-J (3-23) 10 4.5 9.9 λ2-F (2-11) 10 4.9 10.0DH428 3-J (3-23) 10 2.4 5.9 λ2-F (2-11) 10 3.0 7.8

D

Figure 3. Isolation of Antibodies that Specif-

ically Target the CAP206 6-Month Env

(A) PBMCs frommacaque 5173 at week 38 after the

seventh immunization were used for sorting CD20+

CD27+ memory B cells that specifically reacted

with the 6-month Env and not the 30-month Env.

(B) Binding of isolated mAbs DH427 and DH428 to

the seven CAP206 Env immunogen proteins by

ELISA.

(C) Cross-competition of sCD4-Ig binding to the

CAP206 6-month Env by DH427, DH428, and RSV

antibody Synagis determined by ELISA.

(D) Genomic characteristics of DH427 and DH428.

See also Figure S4.

and three (5165, 5167, and 5183) had no autologous nAb res-

ponse, yet all had received the same vaccine immunogen.

This variability in the response suggested that host factors

limited high levels of responses in the other five animals. To

examine the role of cross-reactivity with host antigens for the

development of Tier-2 nAbs, we tested each immunized ma-

caque’s plasma pre- and post-immunization for reactivity to

common auto-antigens recognized by individuals with autoim-

mune disease (Figure S7D). We did not detect significant anti-

body polyreactivity in plasma for any of the auto-antigens

tested.

Next, we tested the DH427 lineage for reactivity with the

same auto-antigens and did not detect reactivity (Figure S7E).

We then tested DH427 and the DH427 unmutated common

ancestor antibodies for reactivity to >9,400 human proteins

on a microarray (Figure 5F). We found that neither DH427

nor the unmutated common ancestor were polyreactive, but

the unmutated common ancestor did react strongly with a sin-

gle host protein Family of Sequence Similarity 21, Member C

(FAM21C; Figure 5F; Table S1). FAM21C is a member of the

WASH complex that is involved in intracellular trafficking (Go-

mez and Billadeau, 2009). These data raise the hypothesis that

reactivity with auto-antigens may have blocked the initial

maturation and/or expansion of DH427-like lineages in the

other five macaques while having the correct germline VL

allele.

Cell Reports 14, 43–

Structure of DH427 in Complex withHIV-1 gp120The DH427/DH428 vaccine-induced line-

age blocked CD4 binding and neutralized

an autologous Tier-2 virus, but its genetic

characteristics differed from those of

most CD4bs nAbs. We crystallized and

determined at 6.6 A resolution the struc-

ture of the DH427 Fab in complex with a

recombinant variant of the ZM1766.66

gp120 that had the CAP206 6-month V5-

loop sequence. We also determined the

structures of the DH427 and DH428 Fabs

at 2.66 A and 2.32 A resolution, respec-

tively (Table S2).

The DH427 Fab associates with HIV

gp120 along an outer rim of the CD4-bind-

ing surface. CDRH2 and CDRH3 contact the V5 loop, and CDRL1

and CDRL3 contact loop E (Figure 6A). The limited resolution did

not allow detailed description of side-chain interactions. We su-

perposed onto the gp120 core domain from our complex the

structure of the gp120 core domain bound with soluble CD4.

The bulk of the Fab was large enough to block CD4 binding,

even though their overlapping footprints differed substantially

(Figure 6B). In contrast, the CD4bs bnAb VRC01 contacts

conserved residues in the V5 loop and has more contacts in the

CD4bs than DH427, allowing for more-effective CD4 blocking

and broad antigen recognition (Figure 6B). Mutability of V5, at

the core of their contact, and a very frequently foundglycosylation

site at position 355 on loop E clearly accounted for the restricted

breadth of DH427andDH428. For example, the glycan at position

463would project directly into the face of the antibody (Figure 6C).

Thus, further affinity maturation would be unlikely to generate

much-greater neutralization breadth.

Superposition of the gp120 outer domain from its complex

with DH427 onto the homologous domain in the BG505 SOSIP

gp140 structure shows that the Fab would project laterally

from the periphery of the trimer (Figure 6D). DH427 is in almost

the same orientation as the CD4bs bnAb VRC01 but displaced

outward by about 25 A. This angle is also distinct from the proto-

typical Tier-1-neutralizing CD4bs mAb F105, which projects in a

more-axial direction, making it prone to steric hindrance on the

‘‘closed’’ trimers of Tier-2 viruses (Figure 6D).

54, January 5, 2016 ª2016 The Authors 47

AVirus DH427 DH428

CAP206 6mo 0.02 0.02CAP206 6mo

S463N >50 >50

Virus DH427 DH428CAP206 T/F >50 >50CAP206 T/F.

6moV5 0.04 4.20

LTRDKDPGQENSNTETFRP

LTRDKD-----SNTETFRP

452 470

CAP206 V5 Sequences

T/F

T/F.6moV5

B C

Figure 4. DH427 and DH428 Neutralize the

Tier-2 CAP206 6-Month Virus

(A) Neutralization of the CAP206 6-month virus and

engineered glycosylation site mutant measured as

the antibody concentration at which RLUs were

reduced 50% compared to control wells (IC50) in the

TZM-bl neutralization assay (mg/ml).

(B) CAP206 T/F V5 sequence and mutated V5

sequence to match the 6-month Env. Red amino

acids were deleted.

(C) DH427 and DH428 neutralization of the T/F and

T/F mutant that had a V5 that matched the 6-month

Env in the TZM-bl neutralization assay (mg/ml; or-

ange, IC50 < 5.0; red, IC50 < 1.0).

See also Figure S5.

DISCUSSION

Recent studies show that bnAb lineages can evolve out of autol-

ogous nAb lineages present early in HIV-1 infection (Doria-Rose

et al., 2014; Liao et al., 2013a; Wibmer et al., 2013). All HIV-1-in-

fected individuals produce autologous Tier-2 nAbs months after

transmission, but not all HIV-infected individuals go on to make

bnAbs (Bar et al., 2012; Doria-Rose et al., 2010; Gray et al.,

2009b; Kwong et al., 2013; Richman et al., 2003; Wei et al.,

2003). Thus, eliciting strain-specific nAbs that target neutraliza-

tion-resistant Tier-2 viruses can be a first step to develop bnAbs

during vaccination. CAP206 is an HIV-infected individual that

produced bnAbs targeting the MPER, but we found that the V5

loop was a target for autologous nAbs early after infection in

CAP206. Thus, in CAP206, the autologous nAb response was

unrelated to the bnAb response. Our finding that, after only

two immunizations with a swarm of CAP206 Envs, the CAP206

6-month variant of the TF Env induced in one of six rhesus ma-

caques potent autologous Tier-2 nAbs that targeted a similar

nAb epitope observed in CAP206 shows that vaccination with

gp140 oligomers can, in principle, also produce such antibodies.

However, that only three macaques had autologous nAbs, and

only one with high titers, contrasts with HIV infection, in which

virtually all HIV-infected individuals make autologous nAbs.

These studies raise several important issues for HIV vaccine

development.

First, our structural studies of DH427 CD4bs autologous nAb

demonstrated that DH427 neutralization was restricted to the

autologous virus because antibody recognition centers on the

V5 loop. Moreover, the mode of V5 contact prevented full

engagement of the CD4bs and explains why this lineage is un-

likely to achieve bnAb breadth with further affinity maturation.

Second, it is a conundrum that all HIV-infected individuals

make autologous nAbs and up to 50% make some level of

bnAbs, whereas no humans vaccinated with Envs have to date

made either autologous nAbs or bnAbs. All the macaques in

the group studied had the IGLV2-F germline allelic variant, and

thus, all had the required VL repertoire to respond to Env immu-

nization. One hypothesis is that there may be host control mech-

anisms for autologous nAbs in both macaques and humans and

that the responding macaque (5173) was permissive for release

of antibodies that can recognize a closed or Tier-2 envelope.

There was no reactivity with auto-antigens common in autoim-

mune disease by the plasma or DH427 lineage antibodies, but

48 Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors

the DH427 unmutated common ancestor antibody did react

with host protein FAM21C. This reactivity may have limited

expansion of this lineage in the other animals.

Broad neutralization appears only after years of persistent

infection, in part because breadth requires high levels of SHM,

even for those antibodies with relatively invariant Env contacts

(Kepler et al., 2014; Kwong and Mascola, 2012; Scheid et al.,

2009, 2011). Driving high SHM by vaccination is particularly diffi-

cult, because many highly mutated antibodies are disfavored by

the host immune system (Haynes and Verkoczy, 2014). DH427

lineage antibodies required SHM for neutralization, but they

were less than 5% mutated and were induced rapidly after

only two immunizations. Thus, the DH427 lineage achieved

potent Tier-2 autologous neutralization with minimal SHM.

Finally, HIV Env shifts from a ‘‘closed’’ to an ‘‘open’’ conforma-

tion when CD4 binds (Liu et al., 2008). In that transition, it ex-

poses epitopes concealed in the closed state. Any particular

Env is in equilibrium between closed and open (and probably

one or more intermediate) states (Julien et al., 2013; Pugach

et al., 2015). The key structural correlate of the distinction be-

tween Tier-1 and Tier-2 is whether this equilibrium is strongly

on the side of the closed conformation (Tier-2, primarily bnAb

site accessible) or more frequently toward the open conforma-

tion (strain-specific and non-neutralizing sites accessible). A

recent study demonstrated that a stabilized recombinant

native-like closed trimer Env (BG505 SOSIP.664) elicited nAbs

to the sequence-matched Tier-2 virus in rabbits and in three of

four macaques (Sanders et al., 2015). In the Sanders et al. study

in macaques, the titers of autologous nAbs ranged from 32 to

168, similar to the titers observed in an earlier study by Willey

et al. (40–113) and in the present study (25–437).The immunogen

used here was a trimeric Env, but not a variant (such as BG505

SOSIP.664) locked into a closed state by intentional modifica-

tion. Nonetheless, it elicited potent autologous nAbs, and in at

least the one lineage we have studied, those nAbs recognized

a site accessible of the surface of a Tier-2 virus and thus, by infer-

ence, on the surface of a closed trimer. Although the DH427 nAb

Env epitope appears to be too variable, both in the length of the

loops and in their potential for glycosylation, to be a target for

bnAbs, its proximity to the CD4bs and its dependence on similar

conformational parameters suggest that related vaccine

immunogens could also induce the germline precursors of

CD4bs bnAbs. Indeed, whereas most Env immunogens fail to

engage germline-reverted CD4bs-directed bnAbs, including

A B

C

D E

F

(legend on next page)

Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors 49

Figure 6. Structure of DH427 in Complex

with HIV gp120

(A) DH427 contact sites within the V5 loop and loop

E (DH427, orange; gp120, gray; V5 loop, red; loop

E, purple; PDB 5F6J).

(B) DH427 in complex with gp120 core with VRC01

(PDB 3NGB) and CD4 (PDB 1GC1) molecule su-

perimposed (DH427, orange; gp120, gray; V5 loop,

red; VRC01, pink; CD4 molecule, yellow).

(C) Co-crystal structure of DH427 and HIV Env

with N-acetylglucosamine (GlcNAc) displayed as

spheres at the V5 glycan site that would interfere

with DH427 heavy chain.

(D)ModelingofDH427,VRC01,andF105 (PDB3HI1)

onto the high-resolution cryo-EM trimer structure of

BG505 SOSIP (BG505 SOSIP trimer [PDB 3J5M],

gray) side view (top) and top view (bottom).

See also Table S2.

germline-reverted VRC01 (McGuire et al., 2013; Zhou et al.,

2010), eliminating glycosylation sites in loop D and V5, as

in the CAP206 6-month variant that induced the DH427 lineage,

allowedDH427 germline-Env binding and B cell lineage develop-

ment. Incorporating these sites, along with truncation or deletion

of V5, into future Env immunogensmay have a useful role in prim-

Figure 5. Reconstruction of the DH427 Clonal Lineage

(A) Binding of DH427, DH428, and the inferred intermediate (I1) and unmutated common ancestor (UCA) t

tibodies were immobilized on a CM5 chip, and protein was flowed over at 100 mg/ml.

(B) Binding of the DH427 I1 and I1-UCA hybrids to the CAP206 6-month Env measured by ELISA.

(C) Amino acid alignment of the germline IGVL2-F*01 with DH427 UCA and I1 antibodies comparedwith IGVL

genomic DNA from animal 5173.

(D) Binding of newly designed DH427 UCA (DH427 UCA2) to the CAP206 6-month Env by SPR. DH427 UCA

flowed over at the indicated concentrations.

(E) Neutralization of DH427 I1, UCA2, and chimeric (UCA VH:I1VL) of the CAP206 6-month virus and glycos

centration (mg/ml) at which RLUs were reduced 50% compared to control wells (IC50) in the TZM-bl neutral

(F) Reactivity of DH427UCA2 and DH427 on a panel of 9,400 human proteins tested using a ProtoArray micro

values are relative fluorescent signal intensity. Each dot represents an average of duplicate array proteins. Da

as cutoff for reactivity.

See also Figures S6 and S7 and Table S1.

50 Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors

ing B cell responses to the CD4bs. More-

over, modifying the CAP206 Env to have a

more-closed structure by SOSIP modifi-

cation may also increase immunogenicity.

Such structural considerations coupled

with understanding host control of induc-

tion of nAbs should lead to design of stra-

tegies for inducing autologous nAbs to

Tier-2 HIV with greater heterologous

neutralization breadth (Liao et al., 2013a).

EXPERIMENTAL PROCEDURES

Donor Subject

CAP206 is an HIV-1 subtype C chronically infected

individual used for this study. This participant was

part of the CAPRISA 002 Acute infection cohort

whose antibody neutralization profile has been

studied since the point of seroconversion (Gray

et al., 2007, 2009b). This study was approved by

the IRBs of the Universities of KwaZulu Natal and

Witwatersrand in South Africa and Duke University. Written informed consent

was obtained from all study participants.

Design andProduction of Recombinant gp140ProteinsDerived from

CAP206

The cloning and sequencing of env sequences was performed as described

previously from longitudinal samples from CAP206 (L.M., unpublished data;

o the CAP206 6-month Env measure by SPR. An-

2-F*02 allele discovered by sequencing the IGVL2-F

2 was immobilized on a CM5 chip, and protein was

ylation site mutant measured as the antibody con-

ization assay.

chip compared with a nonreactive rhesus mAb. Axis

shed lines indicate 500-fold signal/background ratio

Gray et al., 2007; Keele et al., 2008). The single TF and six mutant clones from

later time points (CAP206 TF, 2 months, 6 months, 12 months, 21 months,

24 months, and 30 months) were selected for production as soluble gp140

trimeric immunogens. The recombinant Envs were expressed in 293T cells

and purified using lectin and size-exclusion chromatography as described pre-

viously (Liao et al., 2013b).

Immunization of Rhesus Macaques

Six rhesus macaques each received seven intramuscular immunizations at 6-

week intervals with seven CAP206 Env immunogens (100 mg total) in MF59

adjuvant (Novartis). Blood samples were collected 2 weeks after each immu-

nization. All rhesus macaques were housed at Bioqual. All rhesus macaques

were maintained in accordance with the Association for Assessment and

Accreditation of Laboratory Animals with the approval of the Animal Care

and Use Committees of the NIH and Harvard Medical School. Research was

conducted in compliance with the Animal Welfare Act and other federal stat-

utes and regulations relating to animals and experiments involving animals

and adheres to principles stated in the Guide for the Care and Use of Labora-

tory Animals, NRC Publication, 2011 edition.

Antibody Isolation

CAP206 6-month gp140 and CAP206 30-month gp140 proteins were fluores-

cently labeled with AF647 and BV421 (Invitrogen), respectively. Peripheral

blood mononuclear cells (PBMCs) from rhesus monkey 5173 at week 38

were stained with fluorescently labeled antibodies for cell surface markers

and both CAP206 Envs. Memory B cells that were stained positive for the

CAP206 6-month Env and not the CAP206 30-month Env were sorted into sin-

gle wells of 96-well PCR plates containing RT-PCR buffer as previously

described (Wiehe et al., 2014). Antibody variable heavy and variable light

genes were amplified using nested PCR and purified and sequenced as

described previously (Wiehe et al., 2014). VDJ arrangements, clonal related-

ness, and identification of the intermediate and unmutated common ancestor

were inferred using previously described computational methods (Kepler,

2013; Munshaw and Kepler, 2010).

Sequencing of Germline Variable Region

Genomic DNA was isolated from all six animals from PBMCs at 2 weeks af-

ter the first immunization (QIAmp DNA Blood mini kit; QIAGEN). IGLV2-F se-

quences were amplified using two independent primer sets. To ensure

amplification of non-rearranged variable sequences, both primer sets

reverse primers aligned to sequences present in the non-coding genomic

DNA downstream the V-recombination site. The forward primer for set 1

resided in the IGV2-F leader sequence and upstream of the leader in set

2. The PCR fragments were cloned into a pcDNA2.1 (TOPO-TA kit; Life

Technologies) and transformed into bacteria for sequencing of individual

colonies.

Expression of Recombinant Antibodies

Transient small-scale expression of antibodies was achieved by assembling

VH, VK, or VL sequences with linear cassettes that contain the CMV pro-

moter, respective Ig constant region, and poly A signal sequence using over-

lapping PCR and transfection into 293T cells as described previously (Liao

et al., 2009). Supernatants were directly used to screen for binding of Env an-

tigens in ELISA.

For production of purified recombinant mAbs, the VH and VL genes from

DH427 lineage antibodies were cloned into expression vectors and expressed

and purified as described previously (Liao et al., 2011). Site-directed mutagen-

esis of antibody geneswas performed using theQuikchange II lighteningmulti-

site-directed mutagenesis kit (Agilent).

ELISA binding of transiently transfected supernatants and purified re-

combinant mAbs was performed as described previously (Liao et al.,

2011).

Surface Plasmon Resonance

Surface plasmon resonance assays were performed on a BIAcore 4000 in-

strument, and data analysis was performed with BIAevaluation 4.1 soft-

ware (BIAcore). Anti-gp120 mAbs or sCD4 in 10 mM Na-acetate buffer

(pH 4.5) were directly immobilized to CM5 sensor chips using a standard

amine coupling protocol for protein immobilization. Purified CAP206 Env

glycoproteins were flowed over CM5 sensor chips at concentrations of

2–100 mg/ml. Binding of CAP206 envelope proteins was monitored in

real time at 25�C with a continuous flow of PBS (150 mM NaCl, 0.005%

surfactant P20 [pH 7.4]) at 10–30 ml/min (Alam et al., 2011; Gao et al.,

2014).

Neutralization Assays

Neutralization activities of animal plasma and purified antibodies were deter-

mined by the TZM-bl cell-based neutralization assay (Sarzotti-Kelsoe et al.,

2014). Tier phenotyping of the pseudoviruses was assayed by sensitivity to

a pool of HIV-infected serum as described previously (Seaman et al., 2010).

Antibody Reactivity with Host Proteins

The polyreactivity of the DH427 lineage was assayed in the AtheNA multi-lyte

system (Zeus Scientific). The rhesus mAbs were also tested for reactivity with

human host cellular antigens using ProtoArray 5 microchip (Life Technologies)

compared to a rhesus isotype-matched control antibody as described previ-

ously (Yang et al., 2013).

Expression and Purification of Proteins for Crystallization

The heavy- and light-chain variable and constant domains of the DH427 and

DH428 Fabs were cloned into the pVRC-8400 expression vector using Not1

and Nhe1 restriction sites and the tissue plasminogen activator signal

sequence. The C terminus of the heavy-chain constructs contained a non-

cleavable 63 histidine tag. Fabs were expressed and purified as described

previously (Fera et al., 2014).

The codon-optimized synthetic construct of the ZM176.66 HIV-1 subtype C

gp120 containing aa 41–492 (HXB2 numbering)DV123 (core) was produced by

GenScript with an N-terminal 63-histidine tag and inserted into the pVRC-

8400 expression vector as described for Fabs. The V5 loop of this gp120

core was mutated to that of the CAP206 6-month envelope by site-directed

mutagenesis using manufacturer’s protocols (Stratagene) and expressed in

293S GnTi-suspension adapted cells using linear PEI. After 5 days of expres-

sion, supernatants were clarified by centrifugation and passed over galanthus

nivalis lectin resin pre-equilibrated with 13 PBS. Bound protein was washed

with 13 PBS and eluted with 13 PBS and 500 mMmethyl a-D mannopyrano-

side. The glycoprotein was then deglycosylated overnight at 37�C with Endo

H in a buffer of 50 mM sodium acetate (pH 6.0), 5 mM EDTA, 500 mM NaCl,

1 mg/ml leupeptin, and 1 mg/ml aprotonin. Deglycosylated gp120 core was

then purified by gel filtration chromatography in buffer B (2.5 mM Tris [pH

7.5], 350 mM NaCl, and 0.02% sodium azide) using a superdex 200 analytical

column (GE Healthcare). To make the complex for co-crystallization, DH427

and the ZM176.66 gp120 mutant were mixed in a 1.2:1 molar ratio and incu-

bated for at least 1 hr before passing over a superdex 200 preparatory column

(GE Healthcare) in buffer B. Fractions corresponding to the complex were

combined and concentrated to �10 mg/ml for co-crystallization trials.

Crystallization, Structure Determination, and Refinement

All His-tagged Fabswere crystallized at�15mg/ml, and the DH427/ZM176.66

mutant gp120 core complex was crystallized at �10 mg/ml. Crystals were

grown in 96-well format using hanging drop vapor diffusion and appeared after

24–48 hr at 20�C. DH427 crystals were obtained in a condition of 20% polyeth-

ylene glycol (PEG) 2K monomethyl ether (MME), 100 mM sodium citrate (pH

4.0), and 100 mM NaCl; and DH428 crystals were grown over a reservoir of

40% PEG 400 and 100 mM sodium acetate (pH 5.0). Initial crystals of

DH427/gp120 core complex were obtained in a condition of 1.5 M ammonium

sulfate and 100 mM Tris (pH 8.0) and were optimized to 24-well format to

obtain larger crystals. All crystals were harvested and cryoprotected by the

addition of 20%–25% glycerol to the reservoir solution and then flash cooled

in liquid nitrogen. Crystals of complex were cryoprotected with a series of

different cryoprotectants including 15%–30% glycerol, 15%–30% ethylene

glycol, and 1 or 2 M sodium malonate; however, the best diffraction was ob-

tained with 20%–30% glycerol or ethylene glycol as the cryoprotectant.

Diffraction data were obtained at 100 K from beam lines 24-ID-C and 24-

ID-E at the Advanced Photon Source using a single wavelength. Data sets

Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors 51

from individual crystals (two crystals for DH427 and one crystal for DH428)

were processed with HKL2000 (Otwinowski and Minor, 1997) and XDS

(Kabsch, 2010) for DH427 and DH428, respectively. Molecular replacement

calculations for the free Fabs were carried out with PHASER (McCoy, 2007),

using I3.2 from the CH103 lineage (PDB ID 4QHL) as the starting model. The

I3.2 model was separated into its variable and constant domains for the

DH427 and DH428 Fab structure determinations. Crystals of the DH427 and

DH428 Fabs had four and one molecules per asymmetric unit, respectively.

For both Fabs, subsequent refinement steps were carried out with PHENIX

(Adams et al., 2010), and all model modifications were carried out with Coot

(Emsley and Cowtan, 2004). During refinement, maps were generated from

combinations of positional, group B-factor, and TLS (translation/ libration/

screw) refinement algorithms. Secondary-structure restraints were included

at all stages for all Fabs; noncrystallographic symmetry restraints were applied

to the DH427 Fab throughout refinement.

Data from multiple crystals of the DH427/gp120 core complex were pro-

cessed using HKL2000 (Otwinowski and Minor, 1997), and molecular replace-

ment calculations were carried out with PHASER (McCoy, 2007), using the

refined DH427 coordinates and gp120 core from the VRC01/gp120 complex

(PDB ID 4LST) as the starting models. DH427 was separated into its variable

and constant domains for structure determinations. There were twomolecules

per asymmetric unit. The resulting electron density map for the complex was

further improved by solvent flattening, histogram matching, and noncrystallo-

graphic symmetry averaging using the program Parrot (Winn et al., 2011).

Phase combination was disabled in these calculations. After density modifica-

tion, rigid-body refinement was performed using Refmac in Coot.

Structure validations were performed periodically during refinement using

the MolProbity server (Davis et al., 2007). The final refinement statistics are

summarized in Table S1.

Protein Structure Analysis and Graphical Representations

The core domain of the chimeric ZM176.66 gp120 core mutant designed in our

study was superposed on other gp120 cores from the PDB by least-squares

fitting in Coot. All graphical representations with protein crystal structures

were made using PyMol.

ACCESSION NUMBERS

The accession numbers for the rhesus antibodies reported in this study are

GenBank: KU216183–KU216190. The accession numbers for the structures

of DH427, DH428, and DH427-Env complex reported in this paper are PDB:

5F6H, PDB: 5F6I, and PDB: 5F6J, respectively.

SUPPLEMENTAL INFORMATION

Supplemental Information includes seven figures and two tables and can be

foundwith this article online at http://dx.doi.org/10.1016/j.celrep.2015.12.017.

AUTHOR CONTRIBUTIONS

T.B. isolated and characterized antibodies, designed assays, analyzed and

interpreted data, and wrote and edited the manuscript. D.F. conducted

structural and sequence analyses, protein expression, data analyses and

interpretation, and edited the manuscript. J.B. performed CAP206 plasma

analysis. S.S. assisted with animal study and analysis. K.A. and S.M.A.

performed protein purification and SPR assays. R.Z. and A.F. contributed to

rhesus PCR and antibody production. X.L. assisted with CAP206 Env protein

production. H.-X.L. led protein and antibody production. X.N. and G.K.

performed Protoarrays. K.E.L., C.S., and R.P. performed antibody-binding

assays. L.L.S., R.M.S., andC.M.B. performed rhesus immunizations. L.M. pro-

vided CAP206 Envs and sequences. S.S.A.K. assisted with clinical protocols.

S.B. provided vaccine adjuvant. M.A.M. and D.C.M. performed neutralization

assays. T.B.K. designed software and performed computational analyses of

antibody sequences and inferred unmutated common ancestors. M.A.M. pro-

duced fluorophor-labeled Env proteins for memory B cell staining; S.C.H. led

structural studies and analysis and edited the manuscript; and B.F.H. de-

52 Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors

signed the study, oversaw all experiments, analyzed all data, and wrote and

edited the manuscript.

ACKNOWLEDGMENTS

This work was supported by the Center for HIV/AIDS Vaccine Immunology-

Immunogen Discovery (CHAVI-ID; UMI-AI100645) grant from NIH/NIAID/

DAIDS. D.F. is supported by a F32 fellowship (1F32AI116355-01) from the

NIH. S.C.H. is an investigator with the Howard Hughes Medical Institute. We

thank Dawn J. Marshall and John Whitesides for expert technical assistance

with flow cytometry and Kelly Soderberg and Samantha Bowen for project

management. We also thank beam line staff at Advanced Photon Source

24-ID-C and 24-ID-E for support during data collection.

Received: August 6, 2015

Revised: November 20, 2015

Accepted: December 4, 2015

Published: December 24, 2015

REFERENCES

Adams, P.D., Afonine, P.V., Bunkoczi, G., Chen, V.B., Davis, I.W., Echols, N.,

Headd, J.J., Hung, L.W., Kapral, G.J., Grosse-Kunstleve, R.W., et al. (2010).

PHENIX: a comprehensive Python-based system for macromolecular struc-

ture solution. Acta Crystallogr. D Biol. Crystallogr. 66, 213–221.

Alam, S.M., Liao, H.X., Dennison, S.M., Jaeger, F., Parks, R., Anasti, K.,

Foulger, A., Donathan, M., Lucas, J., Verkoczy, L., et al. (2011). Differential

reactivity of germ line allelic variants of a broadly neutralizing HIV-1 antibody

to a gp41 fusion intermediate conformation. J. Virol. 85, 11725–11731.

Ariyoshi, K., Harwood, E., Chiengsong-Popov, R., and Weber, J. (1992). Is

clearance of HIV-1 viraemia at seroconversion mediated by neutralising anti-

bodies? Lancet 340, 1257–1258.

Bar, K.J., Tsao, C.Y., Iyer, S.S., Decker, J.M., Yang, Y., Bonsignori, M., Chen,

X., Hwang, K.K., Montefiori, D.C., Liao, H.X., et al. (2012). Early low-titer

neutralizing antibodies impede HIV-1 replication and select for virus escape.

PLoS Pathog. 8, e1002721.

Burton, D.R., Poignard, P., Stanfield, R.L., and Wilson, I.A. (2012). Broadly

neutralizing antibodies present new prospects to counter highly antigenically

diverse viruses. Science 337, 183–186.

Davis, I.W., Leaver-Fay, A., Chen, V.B., Block, J.N., Kapral, G.J., Wang, X.,

Murray, L.W., Arendall, W.B., 3rd, Snoeyink, J., Richardson, J.S., and Richard-

son, D.C. (2007). MolProbity: all-atom contacts and structure validation for

proteins and nucleic acids. Nucleic Acids Res. 35, W375–W383.

Doria-Rose, N.A., Klein, R.M., Daniels, M.G., O’Dell, S., Nason, M., Lapedes,

A., Bhattacharya, T., Migueles, S.A., Wyatt, R.T., Korber, B.T., et al. (2010).

Breadth of human immunodeficiency virus-specific neutralizing activity in

sera: clustering analysis and association with clinical variables. J. Virol. 84,

1631–1636.

Doria-Rose, N.A., Schramm, C.A., Gorman, J., Moore, P.L., Bhiman, J.N.,

DeKosky, B.J., Ernandes, M.J., Georgiev, I.S., Kim, H.J., Pancera, M., et al.;

NISC Comparative Sequencing Program (2014). Developmental pathway for

potent V1V2-directed HIV-neutralizing antibodies. Nature 509, 55–62.

Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular

graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132.

Fera, D., Schmidt, A.G., Haynes, B.F., Gao, F., Liao, H.X., Kepler, T.B., and

Harrison, S.C. (2014). Affinity maturation in an HIV broadly neutralizing B-cell

lineage through reorientation of variable domains. Proc. Natl. Acad. Sci.

USA 111, 10275–10280.

Gao, F., Bonsignori, M., Liao, H.X., Kumar, A., Xia, S.M., Lu, X., Cai, F., Hwang,

K.K., Song, H., Zhou, T., et al. (2014). Cooperation of B cell lineages in induc-

tion of HIV-1-broadly neutralizing antibodies. Cell 158, 481–491.

Gomez, T.S., and Billadeau, D.D. (2009). A FAM21-containing WASH complex

regulates retromer-dependent sorting. Dev. Cell 17, 699–711.

Gray, E.S., Moore, P.L., Choge, I.A., Decker, J.M., Bibollet-Ruche, F., Li, H.,

Leseka, N., Treurnicht, F., Mlisana, K., Shaw, G.M., et al.; CAPRISA 002 Study

Team (2007). Neutralizing antibody responses in acute human immunodefi-

ciency virus type 1 subtype C infection. J. Virol. 81, 6187–6196.

Gray, E.S., Madiga, M.C., Moore, P.L., Mlisana, K., Abdool Karim, S.S., Binley,

J.M., Shaw, G.M., Mascola, J.R., and Morris, L. (2009a). Broad neutralization

of human immunodeficiency virus type 1 mediated by plasma antibodies

against the gp41 membrane proximal external region. J. Virol. 83, 11265–

11274.

Gray, E.S., Taylor, N., Wycuff, D., Moore, P.L., Tomaras, G.D., Wibmer, C.K.,

Puren, A., DeCamp, A., Gilbert, P.B., Wood, B., et al. (2009b). Antibody spec-

ificities associated with neutralization breadth in plasma from human immuno-

deficiency virus type 1 subtype C-infected blood donors. J. Virol. 83, 8925–

8937.

Gray, E.S., Madiga, M.C., Hermanus, T., Moore, P.L., Wibmer, C.K., Tumba,

N.L., Werner, L., Mlisana, K., Sibeko, S., Williamson, C., et al.; CAPRISA002

Study Team (2011). The neutralization breadth of HIV-1 develops incrementally

over four years and is associated with CD4+ T cell decline and high viral load

during acute infection. J. Virol. 85, 4828–4840.

Haynes, B.F., and Verkoczy, L. (2014). AIDS/HIV. Host controls of HIV neutral-

izing antibodies. Science 344, 588–589.

Julien, J.P., Cupo, A., Sok, D., Stanfield, R.L., Lyumkis, D., Deller, M.C.,

Klasse, P.J., Burton, D.R., Sanders, R.W., Moore, J.P., et al. (2013). Crystal

structure of a soluble cleaved HIV-1 envelope trimer. Science 342, 1477–1483.

Kabsch, W. (2010). Xds. Acta Crystallogr. D Biol. Crystallogr. 66, 125–132.

Keele, B.F., Giorgi, E.E., Salazar-Gonzalez, J.F., Decker, J.M., Pham, K.T.,

Salazar, M.G., Sun, C., Grayson, T., Wang, S., Li, H., et al. (2008). Identification

and characterization of transmitted and early founder virus envelopes in pri-

mary HIV-1 infection. Proc. Natl. Acad. Sci. USA 105, 7552–7557.

Kepler, T.B. (2013). Reconstructing a B-cell clonal lineage. I. Statistical infer-

ence of unobserved ancestors. F1000Res. 2, 103.

Kepler, T.B., Liao, H.X., Alam, S.M., Bhaskarabhatla, R., Zhang, R., Yandava,

C., Stewart, S., Anasti, K., Kelsoe, G., Parks, R., et al. (2014). Immunoglobulin

gene insertions and deletions in the affinity maturation of HIV-1 broadly reac-

tive neutralizing antibodies. Cell Host Microbe 16, 304–313.

Kwong, P.D., and Mascola, J.R. (2012). Human antibodies that neutralize HIV-

1: identification, structures, and B cell ontogenies. Immunity 37, 412–425.

Kwong, P.D., Mascola, J.R., and Nabel, G.J. (2013). Broadly neutralizing anti-

bodies and the search for an HIV-1 vaccine: the end of the beginning. Nat. Rev.

Immunol. 13, 693–701.

Liao, H.X., Levesque, M.C., Nagel, A., Dixon, A., Zhang, R., Walter, E., Parks,

R., Whitesides, J., Marshall, D.J., Hwang, K.K., et al. (2009). High-throughput

isolation of immunoglobulin genes from single human B cells and expression

as monoclonal antibodies. J. Virol. Methods 158, 171–179.

Liao, H.X., Chen, X., Munshaw, S., Zhang, R., Marshall, D.J., Vandergrift, N.,

Whitesides, J.F., Lu, X., Yu, J.S., Hwang, K.K., et al. (2011). Initial antibodies

binding to HIV-1 gp41 in acutely infected subjects are polyreactive and highly

mutated. J. Exp. Med. 208, 2237–2249.

Liao, H.X., Lynch, R., Zhou, T., Gao, F., Alam, S.M., Boyd, S.D., Fire, A.Z., Ros-

kin, K.M., Schramm, C.A., Zhang, Z., et al.; NISC Comparative Sequencing

Program (2013a). Co-evolution of a broadly neutralizing HIV-1 antibody and

founder virus. Nature 496, 469–476.

Liao, H.X., Tsao, C.Y., Alam, S.M., Muldoon, M., Vandergrift, N., Ma, B.J., Lu,

X., Sutherland, L.L., Scearce, R.M., Bowman, C., et al. (2013b). Antigenicity

and immunogenicity of transmitted/founder, consensus, and chronic envelope

glycoproteins of human immunodeficiency virus type 1. J. Virol. 87, 4185–

4201.

Liu, J., Bartesaghi, A., Borgnia, M.J., Sapiro, G., and Subramaniam, S. (2008).

Molecular architecture of native HIV-1 gp120 trimers. Nature 455, 109–113.

Mascola, J.R., and Haynes, B.F. (2013). HIV-1 neutralizing antibodies: under-

standing nature’s pathways. Immunol. Rev. 254, 225–244.

McCoy, A.J. (2007). Solving structures of protein complexes by molecular

replacement with Phaser. Acta Crystallogr. D Biol. Crystallogr. 63, 32–41.

McGuire, A.T., Hoot, S., Dreyer, A.M., Lippy, A., Stuart, A., Cohen, K.W., Jar-

dine, J., Menis, S., Scheid, J.F., West, A.P., et al. (2013). Engineering HIV

envelope protein to activate germline B cell receptors of broadly neutralizing

anti-CD4 binding site antibodies. J. Exp. Med. 210, 655–663.

Morris, L., Chen, X., Alam, M., Tomaras, G., Zhang, R., Marshall, D.J., Chen,

B., Parks, R., Foulger, A., Jaeger, F., et al. (2011). Isolation of a human anti-

HIV gp41 membrane proximal region neutralizing antibody by antigen-specific

single B cell sorting. PLoS ONE 6, e23532.

Munshaw, S., and Kepler, T.B. (2010). SoDA2: a Hidden Markov Model

approach for identification of immunoglobulin rearrangements. Bioinformatics

26, 867–872.

Otwinowski, Z., and Minor, W. (1997). Processing of X-ray diffraction data

collected in oscillation mode. Methods Enzymol. 276, 307–326.

Pugach, P., Ozorowski, G., Cupo, A., Ringe, R., Yasmeen, A., de Val, N., Derk-

ing, R., Kim, H.J., Korzun, J., Golabek, M., et al. (2015). A native-like

SOSIP.664 trimer based on an HIV-1 subtype B env gene. J. Virol. 89, 3380–

3395.

Richman, D.D., Wrin, T., Little, S.J., and Petropoulos, C.J. (2003). Rapid evo-

lution of the neutralizing antibody response to HIV type 1 infection. Proc. Natl.

Acad. Sci. USA 100, 4144–4149.

Sanders, R.W., van Gils, M.J., Derking, R., Sok, D., Ketas, T.J., Burger, J.A.,

Ozorowski, G., Cupo, A., Simonich, C., Goo, L., et al. (2015). HIV-1 VACCINES.

HIV-1 neutralizing antibodies induced by native-like envelope trimers. Science

349, aac4223.

Sarzotti-Kelsoe, M., Bailer, R.T., Turk, E., Lin, C.L., Bilska, M., Greene, K.M.,

Gao, H., Todd, C.A., Ozaki, D.A., Seaman, M.S., et al. (2014). Optimization

and validation of the TZM-bl assay for standardized assessments of neutral-

izing antibodies against HIV-1. J. Immunol. Methods 409, 131–146.

Scheid, J.F., Mouquet, H., Feldhahn, N., Seaman, M.S., Velinzon, K., Pietzsch,

J., Ott, R.G., Anthony, R.M., Zebroski, H., Hurley, A., et al. (2009). Broad diver-

sity of neutralizing antibodies isolated from memory B cells in HIV-infected in-

dividuals. Nature 458, 636–640.

Scheid, J.F., Mouquet, H., Ueberheide, B., Diskin, R., Klein, F., Oliveira, T.Y.,

Pietzsch, J., Fenyo, D., Abadir, A., Velinzon, K., et al. (2011). Sequence and

structural convergence of broad and potent HIV antibodies that mimic CD4

binding. Science 333, 1633–1637.

Seaman, M.S., Janes, H., Hawkins, N., Grandpre, L.E., Devoy, C., Giri, A., Cof-

fey, R.T., Harris, L.,Wood, B., Daniels, M.G., et al. (2010). Tiered categorization

of a diverse panel of HIV-1 Env pseudoviruses for assessment of neutralizing

antibodies. J. Virol. 84, 1439–1452.

Tomaras, G.D., Binley, J.M., Gray, E.S., Crooks, E.T., Osawa, K., Moore, P.L.,

Tumba, N., Tong, T., Shen, X., Yates, N.L., et al. (2011). Polyclonal B cell re-

sponses to conserved neutralization epitopes in a subset of HIV-1-infected in-

dividuals. J. Virol. 85, 11502–11519.

Walker, L.M., Huber, M., Doores, K.J., Falkowska, E., Pejchal, R., Julien, J.P.,

Wang, S.K., Ramos, A., Chan-Hui, P.Y., Moyle, M., et al.; Protocol G Principal

Investigators (2011). Broad neutralization coverage of HIV by multiple highly

potent antibodies. Nature 477, 466–470.

Wei, X., Decker, J.M., Wang, S., Hui, H., Kappes, J.C., Wu, X., Salazar-Gonza-

lez, J.F., Salazar, M.G., Kilby, J.M., Saag, M.S., et al. (2003). Antibody neutral-

ization and escape by HIV-1. Nature 422, 307–312.

Wibmer, C.K., Bhiman, J.N., Gray, E.S., Tumba, N., Abdool Karim, S.S., Wil-

liamson, C., Morris, L., and Moore, P.L. (2013). Viral escape from HIV-1

neutralizing antibodies drives increased plasma neutralization breadth through

sequential recognition ofmultiple epitopes and immunotypes. PLoS Pathog. 9,

e1003738.

Wiehe, K., Easterhoff, D., Luo, K., Nicely, N.I., Bradley, T., Jaeger, F.H., Den-

nison, S.M., Zhang, R., Lloyd, K.E., Stolarchuk, C., et al. (2014). Antibody light-

chain-restricted recognition of the site of immune pressure in the RV144 HIV-1

vaccine trial is phylogenetically conserved. Immunity 41, 909–918.

Willey, R.L., Byrum, R., Piatak, M., Kim, Y.B., Cho, M.W., Rossio, J.L., Jr.,

Bess, J., Jr., Igarashi, T., Endo, Y., Arthur, L.O., et al. (2003). Control of viremia

and prevention of simian-human immunodeficiency virus-induced disease in

Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors 53

rhesus macaques immunized with recombinant vaccinia viruses plus inacti-

vated simian immunodeficiency virus and human immunodeficiency virus

type 1 particles. J. Virol. 77, 1163–1174.

Winn, M.D., Ballard, C.C., Cowtan, K.D., Dodson, E.J., Emsley, P., Evans,

P.R., Keegan, R.M., Krissinel, E.B., Leslie, A.G., McCoy, A., et al. (2011). Over-

view of the CCP4 suite and current developments. Acta Crystallogr. D Biol.

Crystallogr. 67, 235–242.

Wyatt, R., and Sodroski, J. (1998). The HIV-1 envelope glycoproteins: fuso-

gens, antigens, and immunogens. Science 280, 1884–1888.

54 Cell Reports 14, 43–54, January 5, 2016 ª2016 The Authors

Yang, G., Holl, T.M., Liu, Y., Li, Y., Lu, X., Nicely, N.I., Kepler, T.B., Alam, S.M.,

Liao, H.X., Cain, D.W., et al. (2013). Identification of autoantigens recognized

by the 2F5 and 4E10 broadly neutralizing HIV-1 antibodies. J. Exp. Med.

210, 241–256.

Zhou, T., Xu, L., Dey, B., Hessell, A.J., Van Ryk, D., Xiang, S.H., Yang, X.,

Zhang, M.Y., Zwick, M.B., Arthos, J., et al. (2007). Structural definition of a

conserved neutralization epitope on HIV-1 gp120. Nature 445, 732–737.

Zhou, T., Georgiev, I., Wu, X., Yang, Z.Y., Dai, K., Finzi, A., Kwon, Y.D., Scheid,

J.F., Shi, W., Xu, L., et al. (2010). Structural basis for broad and potent neutral-

ization of HIV-1 by antibody VRC01. Science 329, 811–817.